Self-Similarity in Voice over Internet Traffic

Abstract

Real-time applications, such as Voice over Internet Protocol (VoIP) are now available on many IP networks carriers in the world with lower cost compared to the Public Switched Telephone Network (PSTN). However, current IP networks are based on best-effort service, and do not guarantee the Quality of Service (QoS) of real-time multimedia applications. Time-dependent statistics (e.g., correlation) are important for performance evaluation of IP networks. These statistics can be used to measure the impact of specific impairments. Several studies have found that self-similarity and long-range dependence can have a negative impact in IP traffic, because they give rise to great losses and/or delays. For this reason, it is important to analyze the correlation structures (Short-Range Dependence-SRD and Long-Range Dependence-LRD) of real time applications, such as VoIP. In this talk, it is presented a detailed study of self-similarity behavior in VoIP traffic and its impact in the quality of service.